Evaluation of FORTA Boeing Asphalt Mixture Using Advanced Material Characterization Tests

Transcription

1 Evaluation of FORTA Boeing Asphalt Mixture Using Advanced Material Characterization Tests Prepared by Kamil E. Kaloush, Ph.D., P.E. Assistant Professor Krishna P. Biligiri Maria Carolina Rodezno Luiz de Mello Atish A. Nadkarni Graduate Research Assistants Submitted to FORTA Corporation 100 Forta Drive Grove City, PA May 2007 IRA A. FULTON SCHOOL OF ENGINEERING Department of Civil and Environmental Engineering Tempe, AZ

2 TABLE OF CONTENTS LIST OF TABLES...iii LIST OF FIGURES... iv 1. Introduction Background Study Objective Scope of Work Number of Tests Report Organization MIXTURE CHARACTERISTICS Introduction Separation of Fibers Preparation of Polypropylene Modified FORTA Asphalt Binder BINDER CHARACTERIZATION Introduction Binder Consistency Test Viscosity Temperature Relationship Testing Program Penetration Test Softening Point Test Brookfield TM Viscosity Test Results and Analysis Virgin Binder Modified Binder Comparison of FORTA Binders Comparison of FORTA Modified Binder with other ASU-ADOT Virgin Binders TRIAXIAL SHEAR STRENGTH TEST Background for the Triaxial Shear Strength Test Test Conditions for the Triaxial Shear Strength Test Test Results and Analysis for the Triaxial Shear Strength Test Mohr-Coulomb failure envelope for FORTA Boeing PHX D-1/2 Mixture Influence of Confinement on FORTA Boeing PHX D-1/2 Mix PERMANENT DEFORMATION TEST Background for the Repeated Load Permanent Deformation Test Test Conditions for the Repeated Load Test Repeated Load / Flow Number Test Results and Analysis DYNAMIC MODULUS TEST Introduction Theory of Dynamic Modulus Master Curve Summary of the Test Method Test Data Comparison of FORTA Boeing PHX D-1/2 Mix with ASU-ADOT Asphalt Mixtures Page i

3 7. FATIGUE CRACKING TESTS Background of the Flexural Beam Fatigue Test Testing Equipment Test Procedure and Calculations Materials and Specimen Preparation Testing Factorial Test Results and Analysis SUMMARY AND CONCLUSIONS Summary Conclusions REFERENCES Appendix A Calculation of percentage of fibers in one FORTA bag Calculation of amount of fiber to be added in the Virgin Binder ii

4 LIST OF TABLES Table 1 Mixture Characteristics, FORTA Boeing PHX D-1/ Table 2 Average Aggregate Gradations, FORTA Boeing PHX D-1/ Table 3 Summary of Binder Tests Table 4 Summary of Viscosity-Consistency Tests Results, Virgin Binder Table 5 Summary of Viscosity-Consistency Tests Results, FORTA Modified Binder Table 6 Triaxial Shear Strength Results for FORTA Boeing PHX D-1/2 Mix Table 7 Stress Level / Temperature Combination used for the Repeated Load Test, FORTA Table 8 Master Summary of Repeated Load Test Results, FORTA Boeing PHX D-1/ Table 9 Summary of ε p/ ε r Ratio for Repeated Load Test, FORTA Boeing PHX D-1/ Table 10 Test Conditions of the Dynamic Modulus (E*) Test Table 11 Summary of E* and Phase Angle values for Unconfined FORTA Boeing PHX D-1/2 Mix Table 12 E* Master Curve Parameters of FORTA Boeing PHX D-1/2 Mix Table 13 Summary of Regression Coefficients for the Fatigue Relationships for 70 o F at 50% of Initial Stiffness Table 14 Control Strain Beam Fatigue Test Results for FORTA Boeing PHX D-1/2 Mix Page iii

5 LIST OF FIGURES Figure 1 (a) FORTA Reinforced Fibers, (b) Separated FORTA Fibers Aramid (Yellowcolored) and Polypropylene (Tan-colored)... 2 Figure 2 Advanced Pavements Laboratory, Arizona State University... 2 Figure 3 FORTA Boeing Mix Production at Rinker s Batch Asphalt Plant - Phoenix Figure 4 Gyratory Compaction Equipment, Advanced Pavements Laboratory, ASU... 4 Figure 5 Asphalt Concrete Plug Preparation Procedure... 4 Figure 7 FORTA Modified Asphalt Mix... 8 Figure 6 Preparation of Polypropylene Modified Binder... 9 Figure 8 Viscosity Temperature Relationship of FORTA Binders Figure 9 Comparison of Viscosity Temperature Relationships of FORTA Modified Binder with ASU-ADOT Virgin Binders Figure 10 Comparison of FORTA Boeing PHX D-1/2 Triaxial Strength Test Results at Different Confinement Levels Figure 11 Typical Relationship between Total Cumulative Plastic Strain and Number of Load Cycles Figure 12 Vertical and Radial LVDTs Set Up for an Unconfined Repeated Load Permanent Deformation Test Figure 13 Typical Repeated Load Permanent Strain/Slope Plot, FORTA Boeing PHX D-1/ Figure 14 Comparison of Repeated Load Permanent Deformation Flow Number for the Different Mixtures Figure 15 Comparison of Repeated Load Permanent Deformation % Strain at Failure for Different Mixtures Figure 16 Dynamic (Complex) Modulus Test Figure 17 Specimen Instrumentation of E* Testing Figure 18 FORTA Boeing PHX D-1/2 Mix at Unconfined Condition (a) Construction of Master Curve Average of Three Replicates (b) Master Curve based on Average of Three Replicates (c) Shift Factors based on Average of Three Replicates (d) Master Curve of a Typical Replicate Figure 19 Unconfined Dynamic Modulus Master Curves for FORTA Boeing PHX D-1/2 Mix, Salt River PG and PG ADOT Mixtures Figure 20 Comparison of Measured Dynamic Modulus E* values at 10 Hz for the FORTA Boeing PHX D-1/2 Mix and the Salt River PG and PG Conventional Mixtures at Selected Temperatures Figure 21 Flexural Fatigue Apparatus Figure 22 Loading Characteristics of the Flexural Fatigue Apparatus Figure 23 Manufactured Mold for Beam Compaction Figure 24 Top Loading Platen Figure 25 Specimen sawing Figure 26 Controlled Strain Fatigue Relationships for FORTA Boeing PHX D-1/2 and Salt River Conventional Mixtures Figure 27 Initial Flexural Stiffness Comparison for FORTA Boeing PHX D-1/2 and Salt River Conventional Mixtures Page iv

6 Acknowledgement The authors would like to acknowledge the financial support of FORTA Corporation and TEIJIN TWARON BV to complete this study. Acknowledgements are also due to the Boeing Corporation and Rinker West, Central Region for their assistance in the production and construction of the test section. Special thanks are also due to Mr. Kenny Witczak, Supervisor of the Advanced Pavement Laboratory at ASU for the production and preparation of the laboratory test specimens. v

7 Executive Summary FORTA fibers have been used to improve the performance of asphalt mixtures against permanent deformation and fatigue cracking. Recent development in materials characterization tests in the pavement community necessitated the re-evaluation of the FORTA reinforced asphalt mixtures using state-of-the art testing procedures to demonstrate these performance benefits. A FORTA fiber-reinforced asphalt mixture was placed at the Boeing facility in Mesa, Arizona. The mixture was sampled during the construction project and brought back to the Arizona State University (ASU) laboratories. Mixture preparation included compaction of 150mm diameter gyratory specimens for triaxial testing, and beam specimens prepared and compacted according to AASHTO TP8 test protocols. The target air void level for the test specimens were those typically achieved in the field (about 7%). Rice gravity was determined, and 100mm diameter samples were cored from each gyratory plug prepared. The sample ends were sawed to arrive at typical test specimens of 100mm in diameter and 150mm in height. Thickness and bulk densities were measured in preparation of the testing program. The focus of the laboratory experimental program was on conducting tests that were recommended by the NCHRP 9-19 Project. These tests dealt with recommending Simple Performance Tests (SPT) for the evaluation of asphalt mixtures. The tests included: triaxial shear strength, dynamic (complex) modulus, and repeated load for permanent deformation characterization; flexural beam tests were conducted for fatigue cracking evaluation. The data was used to compare the performance of the Boeing mixture to typical conventional asphalt mixtures. Conventional asphalt binder consistency tests were conducted to develop information that will complement other mixture material properties such as fatigue cracking and permanent deformation. The conventional consistency tests (penetration, softening point and viscosity) were conducted on the FORTA modified binder to determine whether there were any unique characteristics or difficulties in handling the material. The modification process was only done using the Polypropylene fibers. Consistency tests across a wide range of temperatures were conducted according to the accepted American Society for Testing and Materials (ASTM) practices. There were no handling problems or difficulties in adding and mixing the Polypropylene fibers. Based on the test results and analysis, the viscosity-temperature susceptibility relationship at lower temperatures showed no changes from the original virgin binder, which is positive and desirable. At high temperatures, improved properties were observed in having higher viscosities; that is, less susceptible to viscosity change with increased temperatures. Such behavior support the potential performance of the mixture having less low temperature cracking and good resistance to permanent deformation at high temperatures. Triaxial Shear Strength tests were conducted at 100 o F (37.8 o C). These tests provided the standard cohesion and the angle of internal friction parameters of the mixtures. The Mohr- Coulomb failure envelope was developed for the mixture. In addition results from previously tested standard mixtures were included in the analysis in order to compare its properties to those vi

8 obtained for the FORTA Boeing mix. The results of the cohesion parameter showed higher values than normally found for conventional mixtures. Higher cohesion values are indicative of better resistance to shearing stresses. At the same time, the value of angle of internal friction was at the lower end of typical values. The angle of internal friction is an aggregate property that is indicative of the material capacity to develop strength from the applied loads. The post peak failure for the samples tested showed a gradual drop in strength, and it was attributed to the influence of the Aramid fibers. Repeated Load Permanent Deformation tests were conducted and many test parameters were evaluated including tertiary flow (flow number of repetitions). The test results obtained were compared to results available at ASU for conventional mixtures, including test results of mixes that performed very well in national field experiments such as the MnRoad and Westrack. The FORTA Boeing mixture tertiary flow results showed excellent resistance to permanent deformation compared to the other conventional mixtures. Similar to the triaxial shear tests, the post peak (tertiary flow) failure also showed gradual accumulation in permanent strain, a desirable property that was also attributed to the role of the Aramid fibers in the mix. Dynamic Complex Modulus (E*) tests were also conducted and the E* master curve were developed for each mixture. The E* test results for the FORTA Boeing mixture was compared with conventional mixtures test results available from previous studies at Arizona State University. The E* test results at high temperature were 1.5 times higher than those typically found for conventional mixtures, and therefore, the FORTA Boeing mix would provide better performance against permanent deformation. The FORTA Boeing mixture was also subjected to Beam Fatigue Tests. Constant strain fatigue tests were conducted at 70 o F (21 o C) using the beam fatigue apparatus proposed by the Strategic Highway Research Program (SHRP). A comparison was made of the fatigue life obtained for the FORTA Boeing mix with an ADOT PG conventional dense graded mix. The fatigue life was found to be higher for Boeing mix even though both had the same initial flexural stiffness. vii

9 1. Introduction 1.1. Background FORTA fibers have been used to improve the performance of asphalt mixtures against permanent deformation and fatigue cracking. Previous outdated laboratory tests have been conducted to demonstrate performance benefits and to optimize the fiber content in the mixture. Recent development in materials characterization tests in the pavement community necessitated the reevaluation of the FORTA reinforced asphalt mixtures using state-of-the art testing procedures to demonstrate these performance benefits. Figure 1 (a) shows typical FORTA fibers contained in one pound bag. (~445 g). The fibers comprises two types: Aramid (yellow-colored) and Polypropylene (tan-colored). Since 1999, the Department of Civil and Environmental Engineering at Arizona State University (ASU) has been involved with several major asphalt mixtures characterization studies, research being conducted at the Advanced Pavements Laboratory (Figure 2). These studies include the nationally recognized National Cooperative Highway Research Program (NCHRP) 9-19 project (1), which dealt with the development of Simple Performance Tests (SPT) for permanent deformation and cracking potential evaluation of asphalt mixtures. The results from these advanced tests were utilized as input in the newly developed Mechanistic-Empirical Pavement Design Guide (MEPDG) (flexible pavement program developed at ASU). It is noteworthy that ASU has the largest database of HMA Mix engineering properties in the United States, tests conducted on asphalt mixtures from national and international test sites. 1

10 (a) (b) Figure 1 (a) FORTA Reinforced Fibers, (b) Separated FORTA Fibers Aramid (Yellowcolored) and Polypropylene (Tan-colored) Figure 2 Advanced Pavements Laboratory, Arizona State University. 2

11 Furthermore, a long-range asphalt pavement research program is on-going with the Arizona Department of Transportation (ADOT); other studies included work completed for Ford Motor Company, Maricopa County Department of Transportation (MCDOT), Texas DOT and Alberta Transportation, Canada Study Objective The objective of this study was to conduct an advanced laboratory experimental program to obtain typical engineering material properties for FORTA fibers reinforced asphalt mixtures using the most current laboratory tests adopted by the pavement community. The results were compared / ranked among other asphalt mixtures in ASU s database to demonstrate the valueadded uses for asphalt pavement containing FORTA fibers Scope of Work In coordination with FORTA Corporation and Boeing Corporation in Mesa, Arizona, a City of Phoenix asphalt concrete conventional mixture designated as Type D-1/2 single or surface course was selected for paving at the Boeing facility in Mesa. The designated infield section within the construction project used the FORTA fibers as designated and recommended by FORTA Corporation. The adding of fibers was done in coordination and supervision of FORTA representative at Rinker s batch asphalt plant in Phoenix (Figure 3). The mixtures were brought back to the ASU laboratories. Preparation included compaction of 150 mm diameter gyratory specimens for triaxial testing (Figure 4). In addition, beam specimens were prepared and compacted according to AASHTO TP8 test protocols (2, 3). 3

12 Figure 3 FORTA Boeing Mix Production at Rinker s Batch Asphalt Plant - Phoenix. The target air void level for the test specimens was 7% (typical field compaction). Rice gravity was determined. 100 mm diameter samples were cored from each gyratory plug and the sample ends were sawed to get final specimens of 100mm in diameter and 150mm in height (Figure 5). Figure 4 Gyratory Compaction Equipment, Advanced Pavements Laboratory, ASU. Figure 5 Asphalt Concrete Plug Preparation Procedure 4

13 Thickness and bulk densities were measured and the samples were stored in plastic bags in preparation for the testing program. Data obtained from these mixtures were summarized in spreadsheets. The spreadsheet comprised of information such as binder information, aggregates, volumetric mix properties, and the results of the advanced dynamic material characterization tests. The tests included: triaxial shear strength, dynamic (complex) modulus, and repeated load for permanent deformation characterization; and flexural beam tests for fatigue and fracture cracking evaluation. The data were also used to establish a relative ranking of the mixtures according to their expected rutting or cracking potential. Binder tests were conducted to develop information that will complement other mix material properties. The tests provided ASTM A i -VTS i consistency-temperature relationships. These tests were conducted for original conditions and included: penetration, ring and ball softening point, and rotational viscosities at selected temperature range Number of Tests This section summarizes the testing program followed for this study for each mix: Binder Tests - Penetration Test 2 binders x 3 replicates x 5 readings = 30 tests - Softening Point Test 2 binders x 3 replicates = 6 tests - Rotational Viscosity Test 2 binders x 6 temperatures x 1 replicate = 12 tests 5

14 Triaxial Shear Strength 3 confinement levels x 1 temperature x 1 replicate = 3 tests Repeated Load / Flow Number Unconfined x 1 temperature (100 F) x 3 replicates = 3 tests Dynamic Complex Modulus Unconfined x 5 temperatures x 6 frequencies x 3 replicates = 90 Tests Beam Fatigue 1 temperature x 5 strain levels = 5 tests 1.5. Report Organization This report has been divided into eight chapters. Chapter 1 includes the introduction, objective of the study and scope of work. Chapter 2 summarizes the mixture properties, whereas Chapter 3 presents the binder characterization tests. Chapter 4 includes the results for the Triaxial Shear Strength tests; while chapter 5 comprises of permanent deformation test results and analysis. Chapter 6 documents the Dynamic Modulus tests. Fatigue beam tests are included in Chapter 7. Chapter 8 presents the conclusions and recommendations of the study. 6

15 2. MIXTURE CHARACTERISTICS 2.1. Introduction As mentioned earlier, the objective of this study was to conduct a laboratory experimental program to obtain typical engineering material properties for the FORTA fibers Reinforced Boeing PHX D-1/2 asphalt mixture. The reference air voids for the mix was 7.0%. This section provides some information on the mixtures characteristics. The asphalt binder used in the study was PG (4). The hot mix asphalt mixture was obtained as loose mix samples taken from the paver hopper during construction. The mixture properties of the FORTA Boeing PHX D-1/2 project are reported in Table 1 including the maximum theoretical specific gravity that was determined at ASU. Table 2 shows the reported average aggregate gradations for the each mixture (4). Table 1 Mixture Characteristics, FORTA Boeing PHX D-1/2 Binder Mix Design Data Mix Type Binder Type Design AC (%) Target Va (%) Gmm FORTA Boeing PHX D-1/2 PG Table 2 Average Aggregate Gradations, FORTA Boeing PHX D-1/2 Aggregate Gradation Gradation (% Passing) FORTA Boeing PHX D-1/2 2 " " " " " " No No No No Feb 7

16 Figure 6 shows a close up of the FORTA Boeing asphalt mixture that was spread on the table for preparation of the Rice gravity test. Fibers were seen by the naked eye with very good distribution through the out the mix. Figure 6 FORTA Modified Asphalt Mix 2.2. Separation of Fibers As mentioned earlier, FORTA fibers were added to the asphalt mix to obtain an fiber-reinforced asphalt mixture. At ASU, a special task was undertaken to estimate the percentage of each type of fiber. Each bag of fibers comprised of two types: Aramid (yellow-colored) and Polypropylene (tan-colored). Calculation of the percentage of each type of fiber is shown in Appendix A. 8

17 2.3. Preparation of Polypropylene Modified FORTA Asphalt Binder The calculation of the amount of Polypropylene fibers added to the virgin binder is described in Appendix A. The mix time for the preparation was 30 minutes and the mixing temperature ranged between 356 F and 405 F (180 C C). Figure 7 shows the process of the adding the Polypropylene fibers.. Figure 7 Preparation of Polypropylene Modified Binder 9

18 3. BINDER CHARACTERIZATION 3.1. Introduction The objective of binder testing was to compare the effect of soluble polypropylene fibers on the binder, if any. Conventional consistency tests (penetration, softening point and viscosity) were conducted on the extracted binders to determine whether there are any unique characteristics or difficulties in handling the material. Consistency tests across a wide range of temperatures were conducted according to the accepted American Society for Testing and Materials (ASTM) practices. Test results and analysis conducted in this task provided the viscosity-temperature susceptibility of the original (virgin) and polypropylene modified asphalt cement Binder Consistency Test Viscosity Temperature Relationship ASU s experience in regard to applying conventional / standard binder consistency tests to modified asphalt cements had been positive. It has been shown in earlier studies that these test methods can be rational and can be used as a general guide; especially when these consistency tests are being used for descriptive comparative purposes and not for specification control. Most refined asphalt cements, with the exception of heavily air blown or high wax content crudes, exhibit a linear relationship when plotted on a log-log viscosity (centipoises, cp) versus log temperature (in degree Rankine: R = F o F) scale. In this study, centipoise (cp) was selected for this type of plots because the test results are reported in these units. The approach uses only viscosity units (cp) to define the viscosity-temperature relationship. In order to make use of all consistency tests variables over a wide range of temperatures, it was necessary to convert all penetration (pen) and softening point (T RB ) measurements into viscosity units. 10

19 Penetration data was converted to viscosity units by the following model developed at the University of Maryland as a part of a Strategic Highway Research Program (SHRP) study. It should be noted that the following equation is applicable over a very wide range of penetration from 3 to 300. log = log( pen ) (log( pen)) 2 η (3.1) The viscosity obtained from the above equation is in poise (P). The second consistency variable point defined by the softening point (T RB ) is converted to viscosity units by the approach suggested by Shell Oil researchers. It states that all asphalts at their softening point (T RB ) will yield a penetration of approximately 800 and a viscosity of 13,000 poises. The third group of viscosity values at high temperature was obtained by use of the Brookfield Viscometer. Using the above three methods, all penetration and softening point results can be shown or converted to viscosity units, which along with the Brookfield test results can then be used as direct viscosity measurements to obtain a viscosity (η) - temperature (T R ) relationship from the following regression equation: log logη (centipoise) = A + VTS logt (3.2) i i R In Equation (3.2), A i and VTS i represents regression coefficients, which describe the unique consistency-temperature relationship of any blend. The VTS term in this equation represents the slope of the regression equation, which is also interpreted as the Viscosity- Temperature 11

20 Susceptibility parameter. For example, a larger (negative) slope value defines a higher temperature susceptibility of the binder (5, 6) Testing Program The testing program comprised of the tests: Penetration, Ring & Ball, and Brookfield Viscosity. Table 3 summarizes the binder tests conducted at ASU. Table 3 Summary of Binder Tests Age Test Properties Tested Method Test Conditions Tank Penetration Penetration AASHTO T , & 77 F Tank Ring and Ball Softening point AASHTO T53-92 Measured Temp. Tank Brookfield Viscosity Rotational 275, 300, 350, 400, AASHTO TP-48 viscosity 450 & 500 F Penetration Test This test covers the determination of the penetration of semi-solid and solid asphalt binders. The penetration of an asphalt binder is the distance in tenths of a millimeter that a standard penetrates vertically into a sample of the material under fixed conditions of temperature, load and time. This test is commonly used as a measure of consistency. Higher values of penetration indicate softer consistency. The binder sample was heated and cooled under controlled conditions. The penetration was measured with a penetrometer using a standard needle under a specified condition. Penetration tests were conducted at 15 and 25 C (59 and 77 o F) using a 100 g load for 5 second. Penetrations were converted to viscosity using Equation

21 Softening Point Test This test covers the determination of the softening point of asphalt binders using the ring-andball apparatus. Two horizontal disks of binder, cast in shouldered brass rings, are heated at controlled rate in a liquid bath while each supports a steel ball. The softening point is reported as the mean of the temperatures at which the two disks softens enough to allow each ball, enveloped in asphalt binder, to fall a vertical distance of 25 mm. The softening point is used in the classification of asphalt binders and as one of the elements in establishing the uniformity of shipments or sources of supply. The softening point is indicative of the tendency of the binder to flow at elevated temperatures encountered in service. For most asphalt binders, the ring and ball softening point corresponds to a viscosity of 13,000 Poise Brookfield TM Viscosity Test This test determines the viscosity i.e. flow characteristics of asphalt binders at higher temperatures. A Brookfield TM rotational coaxial viscometer was used with a Thermosel TM temperature control system. The rotational viscometer automatically calculates the viscosity at the test temperature. The rotational viscosity is determined by measuring the torque required to maintain a constant rotational speed of a cylindrical spindle while submerged in a binder at a constant temperature. This torque is directly related to the binder viscosity. A rotational viscometer can measure viscosity of asphalt binder both at Newtonian and non-newtonian binder conditions. Unlike capillary tube viscometers, the rotational viscometers have larger clearances between the components and, therefore, are applicable to modified as well as unmodified asphalt binders. The viscosity at different shear rates at different temperatures can be used to determine the viscosity-temperature susceptibility of asphalt binders. 13

22 3.4. Results and Analysis Virgin Binder Table 4 shows results from three consistency tests performed on FORTA Boeing PHX D-1/2 virgin binder. Table 4 Summary of Viscosity-Consistency Tests Results, Virgin Binder Test Temp (C) Temp (F) Temp (R) Log Temp (R) Penetration (.1mm) Viscosity (Poise) Viscosity (cp) Log Log Visc (cp) Penetration E E Penetration E E Softening Point E Brookfield E Brookfield E Brookfield E Brookfield E Brookfield E Brookfield E Temperature - Viscosity Relationship for: FORTA Virgin Binder, Tank Condition 1.0 Log Log (Visc) (cp) y = x R 2 = Log (Temp) (R) Ai = VTSi = R^2 =

23 Modified Binder Table 5 shows results from three consistency tests performed on FORTA Boeing PHX D-1/2 modified binder. Table 5 Summary of Viscosity-Consistency Tests Results, FORTA Modified Binder Test Temp (C) Temp (F) Temp (R) Log Temp (R) Penetration (.1mm) Viscosity (Poise) Viscosity (cp) Log Log Visc (cp) Penetration E E Penetration E E Softening Point E Brookfield E Brookfield E Brookfield E Brookfield E Brookfield E Brookfield E Temperature - Viscosity Relationship for: FORTA Modified Binder, Tank Condition 1.0 Log Log (Visc) (cp) y = x R 2 = Log (Temp) (R) Ai = VTSi = R^2 =

24 3.5. Comparison of FORTA Binders Figure 8 presents the viscosity temperature relationship for the two FORTA binders. The very high coefficients of determinations for both the binders clearly establish the fact that the conducted conventional binder tests were adequate to define the viscosity-temperature susceptibility of the binders. At lower temperatures, the FORTA modified binder has equal or similar viscosity when compared with the virgin binder. At higher temperatures, the FORTA modified binder has higher viscosity, indicating higher stiffness and lower susceptibility to temperature change. 0.9 Modified Log (Log vis) (cp) Virgin y = -3.37x R 2 = y = -3.49x R 2 = Log (Temp) (Rankine) Figure 8 Viscosity Temperature Relationship of FORTA Binders 16

25 3.6. Comparison of FORTA Modified Binder with other ASU-ADOT Virgin Binders Figure 9 shows comparison of the viscosity temperature relationships for the FORTA modified binder with typical ADOT s PG 64-22, PG and PG virgin binders. The below viscosity-temperature curves indicate that the addition of Polypropylene fibers improves the temperature susceptibility of the binder, and improves (increase) the viscosity values at the higher temperatures. At cold temperatures, the FORTA modified binder has a lower viscosity than both PG and PG binders, which is a desirable characteristic. At higher temperatures, the FORTA modified binder has a higher viscosity compared to all the other binders, indicating higher stiffness and lower mix susceptibility to permanent deformation. 1.2 Viscosity FORTA Modified 1.0 y = -3.37x R 2 = 0.99 Log (Log vis) (cp) PG PG PG y = -3.88x R 2 = 0.99 y = -3.84x R 2 = 1.00 y = -3.63x R 2 = Log (Temp) (Rankine) Figure 9 Comparison of Viscosity Temperature Relationships of FORTA Modified Binder with ASU-ADOT Virgin Binders 17

26 4. TRIAXIAL SHEAR STRENGTH TEST 4.1. Background for the Triaxial Shear Strength Test The Triaxial Shear Strength Test has been recognized as the standard test for determining the strength of materials for over 50 years. The results from these tests provide a fundamental basis, which can be employed in analyzing the stability of asphalt mixtures. This is because the stresses acting on the laboratory specimen during the test truly simulate the state of stresses existing in the pavement provided certain specimen boundary and geometry conditions are met. In general, there has been reluctance to adopt this test as a routine test procedure because of the degree of difficulty in performing the test. However, with the improvement in testing equipment and computerized data acquisition systems, an increased interest in the use of the triaxial strength test has been extended to more than just a research tool. The shear strength of an asphalt mixture is developed mainly from two sources: 1) The cementing action of the binder, which is commonly referred to as cohesion from Mohr plots; 2) Strength developed by the aggregate matrix interlock from the applied loads, commonly referred to as φ or the angle of internal friction. The major role and interaction of both of these terms varies substantially with rate of loading, temperature, and the volumetric properties of the mixture. Triaxial tests are run at different confining pressures to obtain the Mohr-Coulomb failure envelope. The Mohr-Coulomb failure envelope is defined by: τ ff = c + σ ff tan φ (4.1) 18

27 where, τ ff = shear stress at failure on failure plane σ ff = normal stress at failure on failure plane c = intercept parameter, cohesion tan φ = slope of the failure envelope ( φ is the angle of internal friction ) Typical c values for conventional AC mixtures are in the range of 5 and 35 psi; whereas typical φ values range between 35 and 48 o. Typical triaxial tests require testing specimens at three or more levels of lateral confinement to accurately develop the failure envelope. Although each test may be run on a single specimen, replicate specimens are desired if higher reliability is required. Specimen size and preparation are also important factors needed to be considered in the testing protocols. Normally, a sample with a height to diameter ratio of 2 is used in order to eliminate the effects of friction against the loading platens and interference of shear cones within the specimen. According to the modified sample preparation protocols used in NCHRP Report 465 (1) (sawed specimen ends and the use of thin lubricated membranes), a sample size of 100 mm (4 inches) in diameter and 150 mm (6 inches) in height was recommended. This size was judged sufficient in providing representative (reproducible) material properties provided the ends are parallel and well lubricated. 19

28 4.2. Test Conditions for the Triaxial Shear Strength Test Three triaxial strength tests, one unconfined and two confined were conducted for the Boeing mixture. These tests provided the standard cohesion c and the angle of internal friction φ parameters. The test was carried out on cylindrical specimens, 100 mm (4 inches) in diameter and 150mm (6 inches) in height, prepared as described previously. The tests were conducted at 100 o F. In addition to the unconfined test, two additional confining pressures were used: 138, and 276 kpa (20 and 40 psi). The specimens were loaded axially to failure, at the selected constant confining pressure, and at a strain rate of 0.05 in/in/min (1.27 mm/mm/min). An IPC Universal Testing Machine (UTM 100) electro- hydraulic system was used to load the specimens. The machine was equipped to apply up to 100 psi (690 kpa) confining pressure and 22,000 lbs (100 kn) maximum vertical load. The load was measured through the load cell, whereas, the deformations were measured through the actuator Linear Variable Differential Transducer (LVDT). Thin and fully lubricated membranes at the sample ends were used to reduce end friction. All tests were conducted within an environmentally controlled chamber throughout the testing sequence, controlled within ±1 o F throughout the entire test Test Results and Analysis for the Triaxial Shear Strength Test The results for the triaxial strength tests for the FORTA Modified mixture are summarized and reported in Table 6. The maximum deviator stress, normal stress and percent strain at failure are summarized for each test condition along with Shear strength parameters, c and φ, as well as failure envelopes. 20

29 Table 6 Triaxial Shear Strength Results for FORTA Boeing PHX D-1/2 Mix. Mix: FORTA Modified Mix - FOB00 Binder: Asphalt Concrete Project: Badger Springs Asphalt Project Air Voids Content: 7% Test Temp: C F Test Machine: UTM100 Binder Content: 5.00% Replicate Data σ3 σ3 (PSI) (KPa) (PSI) Air Voids Sample Strain* Time Sec 0 0 σd FOB σd FOB σd Strain at Maximum Deviator Stress Stress Data FOB Maximum Deviator Stress (PSI) Center σ3 (PSI) σd 1 σd Average Radius Point σ1 (PSI) Cohesion & Angle of Friction Tangent Line Test points C φ Normal Stress (PSI) Shear Stress τ (PSI) Tangent Tangent σ3 Points Average Points Average Results τ = c + σ n tanφ c = 66.1 φ = τ = σ n tan (30) Mohrs Circles Shear Stress(PSI) Mohrs Circles and Individual Tangent Points Normal Stress(PSI) Shear Stress(PSI) Average Tangent Points y = x R 2 = Normal Stress (PSI) 21

30 4.4. Mohr-Coulomb failure envelope for FORTA Boeing PHX D-1/2 Mixture The above table also shows a plot of the Mohr-Coulomb failure envelope represented by the cohesion c and angle of internal friction φ for the tested mixture (1 samples for each confinement: 0, 20 and 40 psi). The parameters c and φ are the strength indicators of the mix. The larger the c value, the larger the mix resistance to shearing stresses. In addition, the larger the value of φ, the larger is the capacity of the asphalt mixture to develop strength from the applied loads, and hence, the smaller the potential for permanent deformation. Though typical c values for conventional AC mixtures have been found to be in the range of 5 and 35 psi, the c value of the FORTA Boeing PHX D-1/2 mix was high, about 66.1; whereas from various research studies, typical φ values have been found to be in the range between 35 and 48 o, but, for the FORTA Boeing PHX D-1/2 mix, it was 30 o. As indicated earlier, φ values are more indicative of the aggregate properties Influence of Confinement on FORTA Boeing PHX D-1/2 Mix Figure 10 presents a comparison plot of the triaxial strength test results for different confinement levels performed on the FORTA Boeing PHX D-1/2 Mix, one replicate per confinement level. The plots represent before and after peak stress development during the test. For the confined tests, it is observed that there is an extension or an endurance of the curve post peak stress, a behavior that was attributed to the influence of Aramid fibers in the FORTA mix. This behavior indicates that Aramid fibers provide additional reinforcement to the asphalt mix in resisting permanent deformation beyond shear failure threshold levels. 22

31 Stress, psi Time, sec 0 psi Conf 20 psi Conf 40 psi Conf Figure 10 Comparison of FORTA Boeing PHX D-1/2 Triaxial Strength Test Results at Different Confinement Levels 23

32 5. PERMANENT DEFORMATION TEST 5.1. Background for the Repeated Load Permanent Deformation Test The repeated load test (Flow Number) is a type of creep used to determine the permanent deformation characteristics of paving materials. For this test, a repeated dynamic load is applied for several thousand repetitions and records the cumulative permanent deformation as a function of the number of cycles (repetitions) over the test period. This approach was employed by Monismith et al. in the mid 1970 s using uniaxial compression tests (7). Several research studies conducted by Witczak et al, used a temperature of 100 or 130 o F, and at 10, 20, or 30 psi unconfined deviator stress level (8). A haversine pulse load of 0.1 sec and 0.9 sec dwell (rest time) is applied for the test duration. A number of parameters describing the accumulated permanent deformation response can be obtained from the test. Figure 11 illustrates the typical relationship between the total cumulative plastic strain and number of load cycles. The cumulative permanent strain curve is generally defined by three zones: primary, secondary, and tertiary. In the primary zone, permanent deformations accumulate rapidly. The incremental permanent deformations decrease reaching a constant value in the secondary zone. Finally, the incremental permanent deformations again increase and permanent deformations accumulate rapidly in the tertiary zone. The starting point, or cycle number, at which tertiary flow occurs, is referred to as the Flow Number (8, 9). 24

33 Secondary Tertiary Primary Figure 11 Typical Relationship between Total Cumulative Plastic Strain and Number of Load Cycles For the analysis of the results, the Francken model developed in 1977 was found to be the most comprehensive representation of the permanent deformation test data. The steps that were used to determine the FN values using this model were as follows (10): Step 1: The mathematical model for the regression analysis was: B DN ε ( N) = AN + C( e 1) (5.1) p Where, ε p (N) is the permanent axial strain (%), N is the number of cycles, and A, B, C and D are regression coefficients. Step 2: Using statistical package software, the model s coefficients were estimated through non linear regression techniques for the test files. 25

34 Step 3: The first derivative of Equation (5.1) with respect to N is found, which gives the strain rate of change as follows: dε p dn = ABN + ( CDe ( B 1) DN ) (5.2) Step 4: The method of estimating the FN for the strain slope curve for this model was done by taking the lowest point of the strain slope against number of cycles. This is where the strain slope starts to increase from a constant value to a higher value. This model was developed based on triaxial repeated load test under various temperatures and stress level. The regression constant C represents whether plastic failure would occur or not Test Conditions for the Repeated Load Test Repeated unconfined load test, were conducted using three replicate specimens of the mixture. All tests were carried out on cylindrical specimens, 100 mm (4 inches) in diameter and 150 mm (6 inches) in height. For the tests, a haversine pulse load of 0.1 sec and 0.9 sec dwell (rest time) was applied. An IPC Universal Testing Machine (UTM 25) electro- pneumatic system was used to load the specimens. The machine is equipped to apply up to 90 psi (620 kpa) confining pressure and 5,500 lb (24.9 kn) maximum vertical load. The load was measured through the load cell, whereas, the deformations were measured through six spring-loaded LVDTs. Two axial LVDTs were mounted vertically on diametrically opposite specimen sides. Parallel studs, mounted on the test specimen, placed 100 mm (4 inches) apart and located at the center of the specimen were used to secure the LVDTs in place. The studs were glued using a commercial 5-minute epoxy. 26

35 An alignment rod with a frictionless bushing was used to keep the studs aligned at extreme failure conditions. Figure 12 shows a photograph of an actual specimen set-up for unconfined test. For radial deformations, four externally mounted LVDTs aligned on diametrical and perpendicular lines were located at the center of the specimen and along opposite specimen sides. The radial LVDTs set-up is also shown in the figure. Thin and fully lubricated membranes at the test specimen ends were used to warrant frictionless surface conditions. The tests were conducted within an environmentally controlled chamber throughout the testing sequence (i.e., temperature was held constant within the chamber to ±1 o F throughout the entire test). The figure shows typical unconfined test set up for the repeated load test. Table 7 summarizes the stress level/temperature combination used for the Repeated Load tests. Figure 12 Vertical and Radial LVDTs Set Up for an Unconfined Repeated Load Permanent Deformation Test 27

36 Table 7 Stress Level / Temperature Combination used for the Repeated Load Test, FORTA Test Type Stress Type * Repeated Load/ Flow Number σ 3 σ d Test Temperature 37.8 o C ( 100 o F) UC (kpa) 0 (psi) 0 (kpa) 207 (psi) 30 * σ 3 confining stress σ d deviator stress UC Unconfined tests 5.3. Repeated Load / Flow Number Test Results and Analysis The results for the repeated load unconfined tests for the FORTA mix are summarized in this section. As it was mentioned before, for the analysis of the test results, the Francken model was used to fit the permanent strain results. This model combines a power model that characterizes the primary and secondary stages of the permanent deformation plot and an exponential model that fits the tertiary stage. The model is differentiated twice and the inflection point is obtained, this point represents the point where tertiary flow occurs (FN). Table 8 includes the flow number, percent of axial strain at failure (Flow), Resilient Modulus at failure, and the permanent deformation parameters (a intercept and b - slope). Table 8 Master Summary of Repeated Load Test Results, FORTA Boeing PHX D-1/2 Mix FORTA Target VA% Temp o F σ 3 (psi) σ d (psi) Axial Flow Number Axial Strain (%) Resilient failure (psi) Intercept a x10-3 (1/psi) Slope b , , , , , ,

37 Table 9 presents the results of the permanent to resilient strain ratio (ε p /ε r ). The MEPDG uses this ratio in the model that predicts permanent deformation in the asphalt layer. Table 9 Summary of ε p/ ε r Ratio for Repeated Load Test, FORTA Boeing PHX D-1/2 Mix FORTA Target VA% Temp o F σ 3 (psi) σ d (psi) Axial Flow Number ε p [%] at Failure ε r [%] at Failure ε p /ε r at Failure , , , Figure 13 represents a typical repeated load/strain plot. Two observed characteristics of this plot is the endurance of the secondary stage and the gradual (less) accumulation of permanent strain beyond tertiary flow. Both were attributed to the presence of the Aramid fibers in the mix, as this behavior is not typically observed in conventional mixes. 1.0 Flow Number Test-FORTA Boeing PHX D-1/2 Temp 100ºF, σ d = 30psi, σ 3 = Accumulated Strain (%) Accumulated Strain Slope(um/m/cycles) Cycles Figure 13 Typical Repeated Load Permanent Strain/Slope Plot, FORTA Boeing PHX D-1/2 29

38 The average Flow Number value of three replicates for the FORTA mix was 46,396 and the average Strain is 0.65%. These results were compared to the results obtain for other mixtures under the same conditions and volumetric properties. Figure 14 and 15 show the comparison of the Flow Number and Strain at failure for the FORTA mix, MnRoad and WesTrack (both mixes represent test sections from national experiments that performed well in the field and under the same laboratory test conditions). Under these conditions, the FORTA mixtures shows higher FN value of the three mixtures and the second highest strain % at failure. 60,000 Repeated Load Permanent Deformation Test Flow Number Temp 100 o F, σ d = 30psi, σ 3 = 0 50,000 46,396 Flow Number (cycles) 40,000 30,000 20,000 10,000 2,041 18,118 0 MnRoad Cell 16 Westrack Cell 19 FORTA Boeing PHX D-1/2 Figure 14 Comparison of Repeated Load Permanent Deformation Flow Number for the Different Mixtures. 30

39 1.00 Repeated Load Permanent Deformation Test Axial Temp 100 o F, σ d = 30 psi, σ 3 = Strain(%) MnRoad Cell 16 Westrack Cell 19 FORTA Boeing PHX D-1/2 Figure 15 Comparison of Repeated Load Permanent Deformation % Strain at Failure for Different Mixtures 31

40 6. DYNAMIC MODULUS TEST 6.1. Introduction The main objective of this section is to summarize test data and master curve parameters obtained from the E* testing and analysis conducted for the FORTA Modified Asphalt Reinforced mix used in the FORTA Boeing Project Theory of Dynamic Modulus For linear viscoelastic materials such as AC mixes, the stress-to-strain relationship under a continuous sinusoidal loading is defined by its complex dynamic modulus (E*). This is a complex number that relates stress to strain for linear viscoelastic materials subjected to continuously applied sinusoidal loading in the frequency domain. The complex modulus is defined as the ratio of the amplitude of the sinusoidal stress (at any given time, t, and angular load frequency, ω), σ = σ 0 sin (ωt) and the amplitude of the sinusoidal strain ε = ε 0 sin(ωt-φ), at the same time and frequency, that results in a steady state response (Figure 16): E* σ σ e iω t 0 = = i ε ε 0e ( ω t φ ) σ 0 sinωt = ε sin( ωt φ) 0 (6.1) Where, σ 0 = peak (maximum) stress ε 0 = peak (maximum) strain φ = phase angle, degrees ω = angular velocity t = time, seconds 32

41 Mathematically, the dynamic modulus is defined as the absolute value of the complex modulus, or: σ o E* = (6.2) ε o Stress-Strain ε o sin(ωt φ) σ o ε o φ / ω Time σ o sin(ωt) Figure 16 Dynamic (Complex) Modulus Test For a pure elastic material, φ = 0, and it is observed that the complex modulus (E*) is equal to the absolute value, or dynamic modulus. For pure viscous materials, φ = 90. The dynamic modulus testing of asphaltic materials is normally conducted using a uniaxially applied sinusoidal stress pattern as shown in the above figure Master Curve In the Mechanistic Empirical Pavement Design Guide (MEPDG), the modulus of the asphalt concrete at all analysis levels of temperature and time rate of load is determined from a master curve constructed at a reference temperature (generally taken as 70 F). Master curves are constructed using the principle of time-temperature superposition. The data at various temperatures are shifted with respect to time until the curves merge into single smooth function. The master curve of the modulus, as a function of time, formed in this manner describes the time dependency of the material. The amount of shifting at each temperature required to form the 33

42 master curve describes the temperature dependency of the material. In general, the master modulus curve can be mathematically modeled by a sigmoidal function described as: Where, Log E* = δ + (6.3) + (log tr ) 1+ e γ β α t r = reduced time of loading at reference temperature δ = minimum value of E* δ+α = maximum value of E* β, γ = parameters describing the shape of the sigmoidal function The shift factor can be shown in the following form: a (T) = t t r (6.4) Where, a (T) = shift factor as a function of temperature t t r T = time of loading at desired temperature = time of loading at reference temperature = temperature While classical viscoelastic fundamentals suggest a linear relationship between log a(t) and T (in degrees Fahrenheit); years of testing by the researchers at ASU have shown that for precision, a second order polynomial relationship between the logarithm of the shift factor i.e. log a(t i ) and the temperature in degrees Fahrenheit (T i ) should be used. The relationship can be expressed as follows: Log a(t i ) = at 2 i + bt i + c (6.5) 34

43 Where, a(t i ) = shift factor as a function of temperature T i T i = temperature of interest, F a, b and c = coefficients of the second order polynomial It should be recognized that if the value of a approaches zero; the shift factor equation collapses to the classic linear form Summary of the Test Method The NCHRP 1-37A Test Method DM-1 was followed for E* testing. For each mix, generally two to three replicates were prepared for testing. For each specimen, E* tests were generally conducted at 14, 40, 70, 100 and 130 F for 25, 10, 5, 1, 0.5 and 0.1 Hz loading frequencies. A 60 second rest period was used between each frequency to allow some specimen recovery before applying the new loading at a lower frequency. Table 10 presents the E* test conditions. Table 10 Test Conditions of the Dynamic Modulus (E*) Test Test Temp. ( F) 14, 40, 70, 100, 130 (Unless otherwise specified) Freq. Rest Period Cycles to Cycles (Hz) (Sec) Compute E* to to to to to to 15 The E* tests were done using a controlled stress mode, which produced strains smaller than 150 micro-strain. This ensured, to the best possible degree, that the response of the material was linear across the temperature used in the study. The dynamic stress levels were 10 to 100 psi for colder temperatures (14 F to 70 F) and 2 to 10 psi for higher temperatures (100 F to 130 F). 35

44 All E* tests were conducted in a temperature-controlled chamber capable of holding temperatures from 3.2 to 140 F ( 16 to 60 C). The mixes were tested in unconfined mode only. The axial deformations of the specimens were measured through two spring-loaded Linear Variable Differential Transducers (LVDTs) placed vertically on diametrically opposite sides of the specimen. Parallel brass studs were used to secure the LVDTs in place. Two pairs of studs were glued on the two opposite cylindrical surfaces of a specimen; each stud in a pair, being 100- mm (4 inch) apart and located at approximately the same distance from the top and bottom of the specimen. Top and bottom surface friction is a very practical problem for compressive type testing. In order to eliminate the possibility of having shear stresses on the specimen ends during testing, pairs of rubber membranes, with vacuum grease within the pairs, were placed on the top and bottom of each specimen during testing. Figure 17 shows the schematic presentation of the instrumentation of the test samples used in the dynamic modulus testing. a. Sample Assembly b. Lateral View Figure 17 Specimen Instrumentation of E* Testing 36

45 6.6. Test Data The quality of the E* test data was checked by Black Space diagrams (E* versus φ), Cole-Cole Plane plots (E* sinφ versus E* cosφ) and E* versus loading frequency plots. Similar to the new MEPDG s input Level-1 approach, E* master curves of all mixtures were constructed for a reference temperature of 70 F using the principle of time-temperature superposition. The timetemperature superposition was done by simultaneously solving for the four coefficients of the sigmoidal function (δ, α, β, and γ) as described in equation 6.3 and the three coefficients of the second order polynomial (a, b, and c) as described in equation 6.5. The Solver function of the Microsoft TM Excel was used to conduct the nonlinear optimization for simultaneously solving these 7 parameters. For the mixture, the set of master curve parameters were obtained for: (i) average E* of all replicates, (ii) E* of all replicates, and (iii) each replicate. The E* of each mix at five test temperatures and six test loading frequencies were also computed using the 7 master curve and shift coefficients. The E* data obtained from both lab and master curve are summarized in Table 11. The Master Curve parameters are summarized in Table 12. Figure 18 shows the construction of Master Curve for the FORTA Boeing PHX D-1/2 mix using the equations 6.3, 6.4 and 6.5, the corresponding shift factors as well as Master curve for a typical replicate. 37

46 Table 11 Summary of E* and Phase Angle values for Unconfined FORTA Boeing PHX D-1/2 Mix Mix Conf. Temp. Freq. Dynamic Modulus, E* (ksi) Phase Angle, φ (degree) (psi) o F Hz Rep1 Rep2 Rep3 Avg. E* %C.V. Rep1 Rep2 Rep3 Avg. φ %C.V. FOB ,788 6,292 5,543 5, ,654 6,211 5,075 4, ,571 6,087 4,911 4, ,385 5,677 4,545 4, ,310 5,426 4,373 4, ,023 4,938 4,047 4, ,889 4,177 5,128 4, ,709 4,091 4,894 4, ,522 3,910 4,600 4, ,150 3,250 4,162 3, ,937 2,998 3,931 3, ,523 2,445 3,420 2, ,457 2,932 4,009 3, ,163 2,663 3,633 2, ,913 2,431 3,243 2, ,418 1,801 2,462 1, ,196 1,610 2,131 1, ,018 1,413 1, ,129 1,399 1,612 1, ,098 1,290 1, ,

47 Table 12 E* Master Curve Parameters of FORTA Boeing PHX D-1/2 Mix Mix Conf (psi) FOB00 0 Parameter Values Based On: MC Parameter Avg. of All data Replicate-1 Replicate-2 Replicate-3 Replicates δ α β γ a b c

48 1.E+07 1.E E* psi 1.E+06 1.E+05 1.E F 40 F 70 F 100 F 130 F Predicted Log Reduced Time, s E* psi 1.E+06 1.E+05 1.E Log Reduced Time, s Predicted E Predicted 1.E Predicted log at 0 E* psi -2 1.E y = -4E-05x x R 2 = Temperature, F 1.E Log Reduced Time, s Figure 18 FORTA Boeing PHX D-1/2 Mix at Unconfined Condition (a) Construction of Master Curve Average of Three Replicates (b) Master Curve based on Average of Three Replicates (c) Shift Factors based on Average of Three Replicates (d) Master Curve of a Typical Replicate 40

49 6.7. Comparison of FORTA Boeing PHX D-1/2 Mix with ASU-ADOT Asphalt Mixtures Two conventional ADOT Salt River mixtures utilizing PG and PG binders were used as a comparison to the FORTA AR mixture. Figure 19 shows the average E* master curves for the two Conventional ADOT mixtures, using unconfined tests, compared with the FORTA modified mixture master curve. The figure can be used for general comparison of the mixtures, but specific temperature-frequency combination values need to be evaluated separately. As it is shown, the Salt River PG mixture shows a higher modulus values at lowest temperatures followed by FORTA mix followed by Salt River PG mix. At room temperature, 21 o C (reduced time = 0 second), FORTA mix is stiffer than the other two while at highest temperatures and lowest frequencies, both FORTA and Salt River PG mixes have similar modulus values followed by the Salt River PG mix. Figure 20 shows a similar, but specific, comparison for selected values of test temperatures (100 and 130 o F) and loading frequency (10 Hz). All the compared three mixes had similar air voids level of 7%. The data was obtained from a previous study conducted at Arizona State University. It is observed that the modulus values are more comparable at higher temperatures: 100 and 130 o F, leading to the observation that the addition of FORTA fibers indeed enhances the properties of an asphalt mixture. In fact, at high temperatures and selected frequency, the FORTA AR mixture had a higher modulus (1.5 times higher) than the other two conventional mixtures; this result supports the potential field performance of the FORTA mix of having better resistance to permanent deformation at high temperatures. 41

50 1.E+07 FORTA Boeing PHX D-1/2 1.E+06 E* psi 1.E+05 Salt River PG Salt River PG E Log Reduced Time, s Figure 19 Unconfined Dynamic Modulus Master Curves for FORTA Boeing PHX D-1/2 Mix, Salt River PG and PG ADOT Mixtures 1,200,000 1,000, , , F 130 F 1,097,269 E*, psi 600, , , , , , ,350 0 Salt River 3/4" PG70-10 Salt River 3/4" PG64-22 FORTA Boeing PHX D-1/2 Figure 20 Comparison of Measured Dynamic Modulus E* values at 10 Hz for the FORTA Boeing PHX D-1/2 Mix and the Salt River PG and PG Conventional Mixtures at Selected Temperatures 42

51 7. FATIGUE CRACKING TESTS 7.1. Background of the Flexural Beam Fatigue Test Load associated fatigue cracking is one of the major distress types occurring in flexible pavement systems. The action of repeated loading caused by traffic induced tensile and shear stresses in the bound layers, which will eventually lead to a loss in the structural integrity of a stabilized layer material. Fatigue initiated cracks at points where critical tensile strains and stresses occur. Additionally, the critical strain is also a function of the stiffness of the mix. Since the stiffness of an asphalt mix in a pavement layered system varies with depth; these changes will eventually effect the location of the critical strain that varies with depth; these changes will eventually effect the location of the critical strain that causes fatigue damage. Once the damage initiates at the critical location, the action of traffic eventually causes these cracks to propagate through the entire bound layer. Over the last 3 to 4 decades of pavement technology, it has been common to assume that fatigue cracking normally initiates at the bottom of the asphalt layer and propagates to the surface (bottom-up cracking). This is due to the bending action of the pavement layer that results in flexural stresses to develop at the bottom of the bound layer. However, numerous recent worldwide studies have also clearly demonstrated that fatigue cracking may also be initiated from the top and propagates down (top-down cracking). This type of fatigue is not as well defined from a mechanistic viewpoint as the more classical bottom-up fatigue. In general, it is hypothesized that critical tensile and/or shear stresses develop at the surface and cause extremely large contact pressures at the tire edges-pavement interface this, coupled with highly aged (stiff) thin surface layer that have become oxidized is felt to be responsible for the surface cracking that 43

52 develops. In order to characterize fatigue in asphalt layers, numerous model forms can be found in the existing literature. The most common model form used to predict the number of load repetitions to fatigue cracking is a function of the tensile strain and mix stiffness (modulus). The basic structure for almost every fatigue model developed and presented in the literature for fatigue characterization is of the following form (11): N f 1 K1 ε t k 2 k 3 1 E k 3 = 2 k = K1 ( ε t ) (E) where: N f = number of repetitions to fatigue cracking ε t = tensile strain at the critical location E = stiffness of the material K 1, K 2, K 3 = laboratory calibration parameters In the laboratory, two types of controlled loading are generally applied for fatigue characterization: constant stress and constant strain. In constant stress testing, the applied stress during the fatigue testing remains constant. As the repetitive load causes damage in the test specimen the strain increases resulting in a lower stiffness with time. In case of constant strain test, the strain remains constant with the number of repetitions. Because of the damage due to repetitive loading, the stress must be reduced resulting in a reduced stiffness as a function of repetitions. The constant stress type of loading is considered applicable to thicker pavement layers usually more than 8 inches. For AC thicknesses between these extremes, fatigue behavior is governed by a mixed mode of loading, mathematically expressed as some model yielding intermediate fatigue prediction to the constant strain and stress conditions. 44

53 7.2. Testing Equipment Flexural fatigue tests are performed according to the AASHTO TP8 (2), and SHRP M-009 (3). The flexural fatigue test has been used by various researchers to evaluate the fatigue performance of pavements (11-14). Figure 21 shows the flexural fatigue apparatus. The device is typically placed inside an environmental chamber to control the temperature during the test. Figure 21 Flexural Fatigue Apparatus The cradle mechanism allows for free translation and rotation of the clamps and provides loading at the third points as shown in Figure 22. Pneumatic actuators at the ends of the beam center it laterally and clamp it. Servomotor driven clamps secure the beam at four points with a predetermined clamping force. Haversine or sinusoidal loading may be applied to the beam via the built-in digital servo-controlled pneumatic actuator. The innovative floating on-specimen transducer measures and controls the true beam deflection irrespective of loading frame compliance. The test is run under either a controlled strain or a controlled stress loading. 45

54 Figure 22 Loading Characteristics of the Flexural Fatigue Apparatus In the constant stress mode, the stress remains constant but the strain increases with the number of load repetitions. In the constant strain test, the strain is kept constant and the stress decreases with the number of load repetitions. In either case, the initial deflection level is adjusted so that the specimen will undergo a minimum of 10,000 load cycles before its stiffness is reduced to 50 percent or less of the initial stiffness. In this study, all tests were conducted in the control strain type of loading Test Procedure and Calculations The test utilized in this study applied repeated third-point loading cycles as was shown in above figure. The sinusoidal load was applied at a frequency of 10 Hz. The maximum tensile stress and maximum tensile strain were calculated as: σ t = P / b h 2 (7.1) ε t = 12 δ h / (3 L 2 4 a 2 ) (7.2) where, σ t = Maximum tensile stress, Pa ε t = Maximum tensile strain, m/m 46

55 P = Applied load, N b = Average specimen width, m h = Average specimen height, m δ = Maximum deflection at the center of the beam, m a = Space between inside clamps, 0.357/3 m (0.119 m) L = Length of beam between outside clamps, m The flexural stiffness was calculated as follow. E = σ t / ε t (7.3) where, E = Flexural stiffness, Pa The phase angle (φ) in degrees was determined as follow. φ = 360 f s (7.4) where, f = Load frequency, Hz s = Time lag between P max and δ max, seconds The dissipated energy per cycle and the cumulative dissipated energy were computed using Equations 6.5 and 6.6, respectively. w = π σ t ε t sin φ (7.5) i N Cumulative Dissipated Energy = = w i (7.6) i= 1 47

56 where, w = Dissipated energy per cycle, J/m 3 w i = w for the i th load cycle During the test the flexural stiffness of the beam specimen was reduced after each load cycle. The stiffness of the beam was plotted against the load cycles; the data was best fitted to an exponential function as follow. E = E i e bn (7.7) where, E = Flexural stiffness after n load cycles, Pa E i = Initial flexural stiffness, Pa e = Natural logarithm to the base e b = Constant N = Number of load cycles Once Equation 6.7 was formulated, the initial stiffness S i can be obtained. Failure was defined as the point at which the specimen stiffness is reduced to 50 percent of the initial stiffness. The number of load cycles at which failure occurred was computed by solving Equation 6.7 for N, or simply: N f,50 = [ln (E f,50 / E i )] / b (7.8) where, N f,50 = Number of load cycles to failure E f,50 = Stiffness at failure, Pa 48

57 7.4. Materials and Specimen Preparation Materials All beam specimens were prepared using the reheated hot mix asphalt FORTA mix that was obtained during construction Mold Assembly The AASHTO TP8-94, and SHRP M-009, flexural fatigue testing protocol, require preparation of oversize beams that later have to be sawed to the required dimensions. The final required dimensions are 15 ± 1/4 in. (380 ± 6 mm) in length, 2 ± 1/4 in. (50 ± 6 mm) in height, and 2.5 ± 1/4 in. (63 ± 6 mm) in width. The procedure does not specify a specific method for preparation. Several methods have been used to prepare beam molds in the laboratory including full scale rolling wheel compaction, miniature rolling wheel compaction, and vibratory loading. In this study beams were prepared using vibratory loading applied by a servo-hydraulic loading machine. A beam mold was manufactured at ASU with structural steel that is not hardened. The mold consists of a cradle and two side plates as shown in Figure 23. The inside dimensions of the mold are 1/2 inch (12 mm) larger than the required dimensions of the beam after sawing in each direction to allow for a 1/4 inch (6 mm) sawing from each face. A top loading platen was originally connected to the loading shaft assembly in the middle as shown in Figure 24. Note that the top platen is made of a series of steel plates welded at the two ends to distribute the load more evenly during compaction. The loading shaft was connected to the upper steel plate rather than extending it to the bottom plate so that an arch effect is introduced that would assist in distributing the load more uniformly. In addition, it was found that if the bottom surface of the 49

58 bottom plate is machined to be slightly concave upward, it would counter balance any bending that might occur during compaction and produce more uniform air void distribution. Figure 23 Manufactured Mold for Beam Compaction Figure 24 Top Loading Platen Specimen Preparation The FORTA mixture was heated at 295 o F (146 o C). The mold was heated separately for one hour at the same temperature as the mix. The mixture was placed in the mold in one load. The mold was then placed on the bottom plate of the loading machine and the top platen was lowered to contact the mixture. 50

59 A small load of 0.2 psi (1.4 kpa) was then applied to seat the specimen. A stress-controlled sinusoidal load was then applied with a frequency of 2 Hz and a peak-to-peak stress of 400 psi (2.8 MPa) for the compaction process. Since the height of the specimen after compaction was fixed, the weight of the mixture required to reach a specified air void value was pre-calculated. Knowing the maximum theoretical specific gravity and the target air voids, the weight of the mixture was determined. During compaction the loading machine was programmed to stop when the required specimen height was reached. After compaction, specimens were left to cool to ambient temperature. The specimens were brought to the required dimensions for fatigue testing by sawing 1/4 inch (6 mm) from each side ( Figure 25). The specimens were cut by using water cooled saw machine to the standard dimension of 2.5 in. (63.5 mm) wide, 2.0 in. (50.8 mm) high, and 15 in. (381 mm) long. Finally, the air void content was measured by using the saturated surface-dry procedure (AASHTO T166, Method A). Figure 25 Specimen sawing 51